url escape : on win and cache

Author: fedemengo

cloudpathlib/s3/s3path.py

@@ -1,5 +1,6 @@
 import os
 import re
+import sys
 from pathlib import Path
 from tempfile import TemporaryDirectory
 from typing import Any, Optional, TYPE_CHECKING
@@ -32,6 +33,7 @@ class S3Path(CloudPath):
     cloud_prefix: str = "s3://"
     client: "S3Client"
     _bucket: str
+    _local_path: Path
 
     @property
     def drive(self) -> str:
@@ -106,3 +108,14 @@ def key(self) -> str:
     @property
     def etag(self):
         return self.client._get_metadata(self).get("etag")
+
+    @property
+    def _local(self) -> Path:
+        if hasattr(self, "_local_path"):
+            return self._local_path
+        no_prefix = self._no_prefix
+        # `:` is invalid in Windows paths; percent-encode it for MRAP ARNs
+        if sys.platform == "win32":
+            no_prefix = no_prefix.replace(":", "%3A")
+        self._local_path = self.client._local_cache_dir / no_prefix
+        return self._local_path

tests/test_s3_specific.py

@@ -394,3 +394,14 @@ def test_mrap_file_operations(s3_rig):
     assert new_file.bucket == _MRAP_ARN
     new_file.unlink()
     assert not new_file.exists()
+
+
+def test_mrap_local_path_windows_encoding(monkeypatch, s3_rig):
+    """On Windows, colons in MRAP ARNs must be percent-encoded in the local cache path."""
+    import cloudpathlib.s3.s3path as s3path_module
+
+    monkeypatch.setattr(s3path_module.sys, "platform", "win32")
+    client = s3_rig.client_class()
+    p = client.CloudPath(f"s3://{_MRAP_ARN}/some/key.txt")
+    assert ":" not in str(p._local), f"Colon found in local path on simulated Windows: {p._local}"
+    assert "%3A" in str(p._local)